Classification and Categorical Inputs with Treed Gaussian Process Models

Recognizing the successes of treed Gaussian process (TGP) models as an interpretable and thrifty model for nonparametric regression, we seek to extend the model to classification. Both treed models and Gaussian processes (GPs) have, separately, enjoyed great success in application to classification problems. An example of the former is Bayesian CART. In the latter, real-valued GP output may be utilized for classification via latent variables, which provide classification rules by means of a softmax function. We formulate a Bayesian model averaging scheme to combine these two models and describe a Monte Carlo method for sampling from the full posterior distribution with joint proposals for the tree topology and the GP parameters corresponding to latent variables at the leaves. We concentrate on efficient sampling of the latent variables, which is important to obtain good mixing in the expanded parameter space. The tree structure is particularly helpful for this task and also for developing an efficient scheme for handling categorical predictors, which commonly arise in classification problems. Our proposed classification TGP (CTGP) methodology is illustrated on a collection of synthetic and real data sets. We assess performance relative to existing methods and thereby show how CTGP is highly flexible, offers tractable inference, produces rules that are easy to interpret, and performs well out of sample.     

MCS: A Method for Finding the Number of Clusters

This paper proposes a maximum clustering similarity (MCS) method for determining the number of clusters in a data set by studying the behavior of similarity indices comparing two (of several) clustering methods. The similarity between the two clusterings is calculated at the same number of clusters, using the indices of Rand (R), Fowlkes and Mallows (FM), and Kulczynski (K) each corrected for chance agreement. The number of clusters at which the index attains its maximum is a candidate for the optimal number of clusters. The proposed method is applied to simulated bivariate normal data, and further extended for use in circular data. Its performance is compared to the criteria discussed in Tibshirani, Walther, and Hastie (2001). The proposed method is not based on any distributional or data assumption which makes it widely applicable to any type of data that can be clustered using at least two clustering algorithms.     

The ribosome uses two active mechanisms to unwind messenger RNA during translation

The ribosome translates the genetic information encoded in messenger RNA into protein. Folded structures in the coding region of an mRNA represent a kinetic barrier that lowers the peptide elongation rate, as the ribosome must disrupt structures it encounters in the mRNA at its entry site to allow translocation to the next codon. Such structures are exploited by the cell to create diverse strategies for translation regulation, such as programmed frameshifting, the modulation of protein expression levels, ribosome localization and co-translational protein folding. Although strand separation activity is inherent to the ribosome, requiring no exogenous helicases, its mechanism is still unknown. Here, using a single-molecule optical tweezers assay on mRNA hairpins, we find that the translation rate of identical codons at the decoding centre is greatly influenced by the GC content of folded structures at the mRNA entry site. Furthermore, force applied to the ends of the hairpin to favour its unfolding significantly speeds translation. Quantitative analysis of the force dependence of its helicase activity reveals that the ribosome, unlike previously studied helicases, uses two distinct active mechanisms to unwind mRNA structure: it destabilizes the helical junction at the mRNA entry site by biasing its thermal fluctuations towards the open state, increasing the probability of the ribosome translocating unhindered; and it mechanically pulls apart the mRNA single strands of the closed junction during the conformational changes that accompany ribosome translocation. The second of these mechanisms ensures a minimal basal rate of translation in the cell; specialized, mechanically stable structures are required to stall the ribosome temporarily. Our results establish a quantitative mechanical basis for understanding the mechanism of regulation of the elongation rate of translation by structured mRNAs.     

A current filamentation mechanism for breaking magnetic field lines during reconnection

During magnetic reconnection, the field lines must break and reconnect to release the energy that drives solar and stellar flares and other explosive events in space and in the laboratory. Exactly how this happens has been unclear, because dissipation is needed to break magnetic field lines and classical collisions are typically weak. Ion-electron drag arising from turbulence, dubbed 'anomalous resistivity', and thermal momentum transport are two mechanisms that have been widely invoked. Measurements of enhanced turbulence near reconnection sites in space and in the laboratory support the anomalous resistivity idea but there has been no demonstration from measurements that this turbulence produces the necessary enhanced drag. Here we report computer simulations that show that neither of the two previously favoured mechanisms controls how magnetic field lines reconnect in the plasmas of greatest interest, those in which the magnetic field dominates the energy budget. Rather, we find that when the current layers that form during magnetic reconnection become too intense, they disintegrate and spread into a complex web of filaments that causes the rate of reconnection to increase abruptly. This filamentary web can be explored in the laboratory or in space with satellites that can measure the resulting electromagnetic turbulence.     

On Definitions of Information in Physics

During the refereeing procedure of Anthropomorphic Quantum Darwinism by Thomas Durt, it became apparent in the dialogue between him and me that the definition of information in Physics is something about which not all authors agreed. This text aims at describing the concepts associated to information that are accepted as the standard in the Physics world community.     

Displacing Epistemology: Being in the Midst of Technoscientific Practice

Interest the Erklären?CVerstehen debate is usually interpreted as primarily epistemological. By raising the possibility that there are fundamentally different methods for fundamentally different types of science, the debate puts into play all the standard issues??that is, issues concerning scientific explanation and justification, the unity and diversity of scientific disciplines, the reality of their subject matter, the accessibility of various subject matters to research, and so on. In this paper, however, I do not focus on any of these specific issues. I start instead from the fact that the very existence of the debate itself is an issue; in fact, it poses a philosophical problem that almost everyone but the hardest line logical empiricists has come to realize cannot be resolved epistemologically. In my view, however, that it cannot be resolved ontologically, either. I think the problem is at bottom hermeneutical, and its resolution requires that we focus first, not on the objects of science or the methods of studying them, but on the character of the philosophical orientation assumed by those who would try to resolve it. In this paper, I explain why I think this is so by analyzing (1) Dilthey??s contribution to the original debate, (2) Husserl??s reaction to Dilthey, and (3) Heidegger??s critical evaluation of both. This line of philosophical development??this movement of self-understanding from critiques of objectivism to hermeneutical phenomenology??is of course already a central feature of much work in continental philosophy of science. In my conclusion, however, I argue for the less well-established??even if apparently approved??idea that it ought to be a central feature of technoscience studies as well.     

Complement factor H binds malondialdehyde epitopes and protects from oxidative stress

Oxidative stress and enhanced lipid peroxidation are linked to many chronic inflammatory diseases, including age-related macular degeneration (AMD). AMD is the leading cause of blindness in Western societies, but its aetiology remains largely unknown. Malondialdehyde (MDA) is a common lipid peroxidation product that accumulates in many pathophysiological processes, including AMD. Here we identify complement factor H (CFH) as a major MDA-binding protein that can block both the uptake of MDA-modified proteins by macrophages and MDA-induced proinflammatory effects in vivo in mice. The CFH polymorphism H402, which is strongly associated with AMD, markedly reduces the ability of CFH to bind MDA, indicating a causal link to disease aetiology. Our findings provide important mechanistic insights into innate immune responses to oxidative stress, which may be exploited in the prevention of and therapy for AMD and other chronic inflammatory diseases.     

Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk

Blood pressure is a heritable trait influenced by several biological pathways and responsive to environmental stimuli. Over one billion people worldwide have hypertension (≥140?mm?Hg systolic blood pressure or ≥90?mm?Hg diastolic blood pressure). Even small increments in blood pressure are associated with an increased risk of cardiovascular events. This genome-wide association study of systolic and diastolic blood pressure, which used a multi-stage design in 200,000 individuals of European descent, identified sixteen novel loci: six of these loci contain genes previously known or suspected to regulate blood pressure (GUCY1A3-GUCY1B3, NPR3-C5orf23, ADM, FURIN-FES, GOSR2, GNAS-EDN3); the other ten provide new clues to blood pressure physiology. A genetic risk score based on 29 genome-wide significant variants was associated with hypertension, left ventricular wall thickness, stroke and coronary artery disease, but not kidney disease or kidney function. We also observed associations with blood pressure in East Asian, South Asian and African ancestry individuals. Our findings provide new insights into the genetics and biology of blood pressure, and suggest potential novel therapeutic pathways for cardiovascular disease prevention.     

Evidence for a spin-aligned neutron-proton paired phase from the level structure of (92)Pd

Shell structure and magic numbers in atomic nuclei were generally explained by pioneering work that introduced a strong spin-orbit interaction to the nuclear shell model potential. However, knowledge of nuclear forces and the mechanisms governing the structure of nuclei, in particular far from stability, is still incomplete. In nuclei with equal neutron and proton numbers (N = Z), enhanced correlations arise between neutrons and protons (two distinct types of fermions) that occupy orbitals with the same quantum numbers. Such correlations have been predicted to favour an unusual type of nuclear superfluidity, termed isoscalar neutron-proton pairing, in addition to normal isovector pairing. Despite many experimental efforts, these predictions have not been confirmed. Here we report the experimental observation of excited states in the N = Z = 46 nucleus (92)Pd. Gamma rays emitted following the (58)Ni((36)Ar,2n)(92)Pd fusion-evaporation reaction were identified using a combination of state-of-the-art high-resolution γ-ray, charged-particle and neutron detector systems. Our results reveal evidence for a spin-aligned, isoscalar neutron-proton coupling scheme, different from the previous prediction. We suggest that this coupling scheme replaces normal superfluidity (characterized by seniority coupling) in the ground and low-lying excited states of the heaviest N = Z nuclei. Such strong, isoscalar neutron-proton correlations would have a considerable impact on the nuclear level structure and possibly influence the dynamics of rapid proton capture in stellar nucleosynthesis.     

Quantum gates and memory using microwave-dressed states

Trapped atomic ions have been used successfully to demonstrate basic elements of universal quantum information processing. Nevertheless, scaling up such methods to achieve large-scale, universal quantum information processing (or more specialized quantum simulations) remains challenging. The use of easily controllable and stable microwave sources, rather than complex laser systems, could remove obstacles to scalability. However, the microwave approach has drawbacks: it involves the use of magnetic-field-sensitive states, which shorten coherence times considerably, and requires large, stable magnetic field gradients. Here we show how to overcome both problems by using stationary atomic quantum states as qubits that are induced by microwave fields (that is, by dressing magnetic-field-sensitive states with microwave fields). This permits fast quantum logic, even in the presence of a small (effective) Lamb-Dicke parameter (and, therefore, moderate magnetic field gradients). We experimentally demonstrate the basic building blocks of this scheme, showing that the dressed states are long lived and that coherence times are increased by more than two orders of magnitude relative to those of bare magnetic-field-sensitive states. This improves the prospects of microwave-driven ion trap quantum information processing, and offers a route to extending coherence times in all systems that suffer from magnetic noise, such as neutral atoms, nitrogen-vacancy centres, quantum dots or circuit quantum electrodynamic systems.